JPS63148115A - Position detection system - Google Patents

Abstract

PURPOSE:To improve the accuracy of position detection by registering the estimated position of a position detection system based upon integral data as its position, and selecting a correlation coefficient indicating that an error about a road is minimum and outputting a corresponding estimated position. CONSTITUTION:A traveling distance detection signal and a traveling direction detection signal which are outputted by a traveling distance detection part 1 and a traveling direction detector 2, map data stored in a road map memory 3, and evaluation data outputted from a correlation coefficient evaluation part 6 are inputted to an estimated position calculation part 4 to integrate an estimated position. A correlation coefficient calculation part inputs the map data stored in the memory 3 and position data calculated by the calculation part 4 to calculate the correlation coefficient of the registered estimated position as to respective roads. An evaluation part 6 selects the correlation coefficient indicating that the error regarding the roads is least and outputs the estimated position corresponding to the selected correlation position as a current position. Thus, the accuracy and likelihood ratio of the position detection are improved.

Description

[Detailed Description of the Invention] <Industrial Application Field> The present invention relates to a position detection method, and more specifically,
The location of vehicles traveling at any point on the road transportation network,
This invention relates to a position detection method that detects data without receiving data from the outside using radio waves or the like.

<Conventional technology> Conventionally, methods for detecting the position of a vehicle traveling at any point on a road transportation network include a distance sensor, a direction sensor, and a process that performs the necessary processing on the output signals from both sensors. Dead Reckoning (Dead Reckoning), which is equipped with a device and obtains current position data of the vehicle while integrating distance changes and azimuth changes that occur as the vehicle travels.
has been proposed, but there is a problem in that the errors that the distance sensor and direction sensor inevitably have accumulate as the vehicle continues to travel, and the errors included in the obtained current position data also accumulate. .

Considering these problems and assuming that the vehicle is traveling on the road, the current position data obtained based on the dead reckoning navigation described above is compared with the road transportation network data stored in the memory in advance. A map/matching method that calculates the amount of deviation of the current position data from the road as a cumulative error, corrects the current position data by the cumulative error amount, and matches the current position data with the road data. It has been proposed (U.S. Pat. No. 3,789.198, JP-A-58-99715, JP-A-18-11)
Publication No. 3711, and “LANDPALL A 11
1GH-RESOLUTION AUTOMATIC CV
EHICLE-LOCATION SYSTEM',
D, KING, GECJournal of'
5science & Technology, Vo
l, 45.

fVll, 1978).

Specifically, ■ In the position detection method described in U.S. Pat. The current position data of the vehicle is calculated based on the data, the calculated current position data is compared with the road position data stored in the memory in advance, and if the difference between both data is within a predetermined threshold, The current position data is corrected so that it is located on the nearest road, and conversely, if the difference between both data is greater than a predetermined threshold, no correction is performed at all, so that the correction is performed with high overall accuracy. , I'm trying to display the exact current location.

■ In the position detection method described in JP-A-58-99715, the current position data is updated by calculating the distance change component in the coordinate axis direction of the road map every time a predetermined distance is traveled. When the vehicle is off the road, the current location is displayed accurately by simply using the location data on the nearest road as the current location data.

■ In the position detection method described in JP-A-58-113711, the error of a direction sensor that detects the direction of travel based on geomagnetism is compared by comparing the curvature obtained based on the direction of travel with the curvature of the road. By doing so, we are trying to correct the current position and display the current position accurately.

■ LANDFALL A H'1GH-RESOLU
TION AUTOMATICVEHICLE-LOC
In the position detection method described in the ATION SYSTEM, the road transportation network is understood as a non-branching section and a plurality of typical branching sections, and the vehicle is traveling from one of the branching sections toward another branching section. The distance to another branch in the case is calculated based on the signal from the distance sensor, and when it is determined that the branch has been reached, the subsequent change in the running direction is detected by the azimuth sensor, By determining which branch exit the vehicle is headed to, the current location of the vehicle is displayed accurately as its location on the road.

<Problems to be Solved by the Invention> Even in the above direction position detection method, when the road transportation network is relatively simple, the increase in cumulative error is corrected and Although it is possible to accurately display the current position, for example, when approaching a truck, etc., or when driving under elevated tracks or at railroad crossings, it is affected by external magnetic fields, so it has a higher error rate than a distance sensor. The cumulative error caused by the direction sensor, which has a large proportion, will be included in the vehicle's current position data as an error that is difficult to recover from, and if the current position data is simply corrected to correspond to the nearest road, the actual However, there is a problem in that the vehicle may be displayed as traveling on a completely different road than the one on which it is currently traveling.

Furthermore, even in distance sensors that detect the distance traveled by a vehicle, the cumulative error may exceed the allowable limit error due to the influence of tire air pressure, etc. In this case as well,
This results in the same problem as in the above case.

Even if the probability of the occurrence of the above problem is low, once it occurs, subsequent vehicle position detection will be based on inaccurate current position data. There is a problem in that all detected vehicle position data is inaccurate, and this method cannot be put to practical use as a position detection method for vehicles.

<Object of the invention> This invention was made in view of the above problems,
It is an object of the present invention to provide a position detection method that can accurately detect the current position of a vehicle without being affected by the complexity of road transportation networks.

<Means for Solving the Problems> In order to achieve the above object, the position detection method of the present invention is based on the self-detection method determined based on the accumulated data obtained by obtaining the distance traveled and the amount of change in direction at each predetermined timing. , and the amount of error in the estimated position determined based on the error between the integrated data and the road map, and calculate the estimated position by corresponding to all roads located within the range of error centered on the estimated position. is registered as a self-position, calculates the correlation coefficient of the registered estimated position for each road, selects the correlation coefficient that shows the least error for the road, and estimates the position corresponding to the selected correlation coefficient. The position is output as the current position.

However, the largest correlation coefficient may be selected and the estimated position corresponding to the largest correlation coefficient may be output as the current position.

In this case, if the correlation coefficient becomes less than a predetermined value, the registration of the corresponding estimated position may be deleted, or if the difference between the correlation coefficients becomes more than a predetermined value. Alternatively, the estimated position corresponding to the smaller correlation coefficient may be deregistered. In these cases, the registration of the estimated position may be deleted after passing through a predetermined unit of the road.

Furthermore, when the estimated position with the largest correlation coefficient and the estimated position with the second largest correlation coefficient are switched, the largest correlation coefficient may be selected with a predetermined hysteresis characteristic. Alternatively, if there are a plurality of estimated positions with the maximum correlation coefficient, the estimated position closest to the center of the road may be output as the current position.

Furthermore, the distance traveled and the amount of change in direction may be obtained and accumulated at predetermined intervals, or the distance traveled and the amount of change in direction may be obtained and accumulated at every predetermined distance. Good too.

With the above position detection method, the distance traveled and the amount of changes in direction are obtained and integrated at predetermined timings, and based on the integrated data, the distance detected on any road on the road map stored in memory is calculated. When detecting one's own position, one obtains one's estimated position based on the integrated data, and also obtains f!
Obtain the limit error amount corresponding to the estimated position determined based on the error between the 215 data and the road map, and calculate the above estimated position automatically by making it correspond to all roads located within the limit error amount centering on the estimated position. Register it as a location, calculate the correlation coefficient for each road of the registered estimated location, select the correlation coefficient that shows the least error for the road,
The estimated position corresponding to the selected correlation coefficient can be output as the current position.

In summary, while monitoring the correlation between the vehicle's estimated position and the road map, at least one estimated position that satisfies a predetermined correlation is registered, and all registered estimated positions are It is possible to calculate fluctuations in the correlation coefficients, select the correlation coefficient showing the smallest error with respect to the road, and output the estimated position corresponding to the selected correlation coefficient as the current position.

Then, if the largest correlation coefficient is selected and the estimated position corresponding to the largest correlation coefficient is output as the current position, the correlation coefficient with the least error relative to the road is the largest correlation coefficient. Therefore, the estimated position corresponding to the largest correlation coefficient can be output as the current position.

In addition, in this case, if the registration of the corresponding estimated position is to be deleted when the correlation coefficient becomes less than a predetermined value, the correlation coefficient becomes less than the predetermined value as the vehicle travels. Since the registration of the estimated position is deleted, the number of saved estimated positions can be reduced, and the final selection of the correlation coefficient can be facilitated. On the other hand, if the estimated position corresponding to the smaller correlation coefficient is deleted when the difference between the correlation coefficients exceeds a predetermined value, the correlation coefficient changes as the vehicle moves. Since the registration of the estimated position with the smaller correlation coefficient is deleted when the difference between them exceeds a predetermined value, the number of saved estimated positions can be reduced, and the final selection of the correlation coefficient can be facilitated.

In these cases, if the estimated position is deregistered after passing through a predetermined unit of the road, the influence of temporary fluctuations in the correlation coefficient can be eliminated.

Furthermore, if the estimated position with the largest correlation coefficient and the estimated position with the second largest correlation coefficient are switched, the largest correlation coefficient is selected with a predetermined hysteresis characteristic. The maximum correlation coefficient is selected based on the magnitude relationship of the correlation coefficients after stabilization, without being affected by the magnitude relationship of local correlation coefficients, and the estimated position corresponding to the selected correlation coefficient is currently calculated. It can be output as a position.

Also, if there are multiple estimated positions with the maximum correlation coefficient, and the estimated position closest to the center of the road is output as the current position, the estimated position closest to the center of the road is output as the current position. be able to.

Furthermore, the distance traveled and the amount of change in direction are calculated and accumulated at predetermined intervals, or the distance traveled and the amount of change in direction are calculated and calculated at each predetermined distance. Also, it is possible to select the correlation coefficient that shows the smallest error with respect to the road based on the integrated data, and output the estimated position corresponding to the selected correlation coefficient as the current position.

<Example> Hereinafter, embodiments will be described in detail with reference to the accompanying drawings showing examples.

FIG. 1 is a block diagram showing an embodiment of a device for implementing the position detection method of the present invention, in which a traveling distance detection section (
1), a driving direction detection unit (2), a road map memory (3) that stores the road transportation network in a predetermined range in advance, and a travel distance detection signal output from the rain detection unit (1) [2]. , an estimated position calculation unit (which calculates an estimated position by inputting the driving direction detection signal, map data stored in the road map memory (3), and evaluation data output from a correlation coefficient evaluation unit (6) to be described later); 4) and the above road map memory (3)
The map data stored in
) and a correlation coefficient calculation unit (5) that calculates a correlation coefficient by inputting device data consisting
) is used as an input to evaluate the magnitude relationship of the correlation coefficients and output the current position. The part [1] has, for example, a photoelectric switch that detects the rotation of the wheel,
The number of rotations of the wheel is obtained by counting the number of output pulse signals from the photoelectric switch with a counter, and the count data output from the counter is multiplied by a predetermined constant indicating the outer circumference of the wheel using a multiplier. Any configuration that calculates the traveling distance per unit time may be used, but any conventionally known configuration may be used, such as one that calculates the traveling speed of the vehicle based on Dotsupura shift, etc., and calculates the traveling distance by integrating it. are available.

As the traveling direction detecting section [a], for example, a geomagnetic sensor that outputs rotation angle data per unit time by detecting the horizontal component of geomagnetism can be used, but a gyro or the like can also be used.

The road map memory (3) is one in which road map data (combination data of points and lines indicating the direction of the road, distance between branch points, etc.) covering a predetermined range is stored in advance, and includes a semiconductor memory, Cassette tapes, CD-ROMs, etc. can be used.

The estimated position calculation section (4) includes the travel distance detection section (1).
), and the rotation angle data Δθ output from the traveling direction detection unit [2], the east-west direction component ΔX (-△JX cos △θ) of the above-mentioned distance is calculated.
, and the north-south direction component △y (-△JXsln△θ), and by adding X and △y to each of the above components to the previous position data (Px', Py'), the current position data is calculated. (PK, Py), and corresponds to the above distance data against the marginal error (including distance error, rotation angle error, and road map error) that the previous position data may have. The limit error that the current position data may have is calculated by adding the increase in limit error. Also, based on the above distance data △'' and rotation angle data △θ, curve
Among the map data read from the road map memory (3), the position data (Px,
Positions on the road that are within the limit error range centered on Py) are registered as estimated positions in a memory (not shown), and all estimated positions are displayed on a display (not shown).

The correlation coefficient calculation unit 10 includes the estimated position calculation unit (4)
The similarity between the movements of estimated positions on all the roads registered in the road map memory (3) and the roads in the road map memory (3) is calculated. To explain in more detail, the correlation coefficient changes as the vehicle travels, and for example, the previous correlation coefficient is 71. j and the correlation coefficient obtained by this calculation is △γj, then the new correlation coefficient γDi, j is γ i + L, j −Ax γ i, j + BX △γj
(However, j is a coefficient indicating the registered estimated position. Also, A and B are weighting coefficients, and if A-B-0,5 is set, it becomes a simple average). Specifically, the correlation coefficient of the estimated position corresponding to the road with the least error relative to the road is the largest. However, by varying the calculation criteria for the correlation coefficient, it is also possible to make the correlation coefficient of the estimated position corresponding to the road with the least error relative to the road the smallest.

The correlation coefficient evaluation unit (6) evaluates the magnitude relationship between the correlation coefficients corresponding to each of the estimated positions, stores the estimated position corresponding to the largest correlation coefficient as the actual position, and saves the estimated position corresponding to the largest correlation coefficient as the actual position. This is to delete the registration of the estimated location. To explain in more detail, the magnitude relationship between the above correlation coefficients is evaluated at predetermined timings, and the registration of the estimated position corresponding to the correlation coefficient that has become smaller than a predetermined threshold is deleted, so that the registration is not deleted. The correlation coefficient corresponding to the estimated position will be sequentially updated to a new correlation coefficient in the correlation coefficient calculation unit [5], and the above evaluation will be performed again. Then, by performing the above series of evaluation operations,
In the end, only one estimated position remains due to the influence of road branches, etc., and this remaining estimated position continues to be displayed as the current position.

The position detection operation by the apparatus having the above configuration will be explained in detail with reference to FIGS. 2 to 5.

First, explanation will be given based on the examples shown in FIGS. 2 and 3.

Figure 2 is a diagram showing a part of the road transportation network.
, L 1. L 2. Only L3 is shown, and the vehicle is traveling from road LO through road LL, as indicated by a broken line S. In addition, E indicates a limit error, a, b, c, d indicate a branch or a curve,
γ0. γl, γ2. γ3 indicates a correlation coefficient corresponding to the estimated position on each road.

Moreover, FIG. 3 is a diagram showing the change in the correlation coefficient γ when the vehicle runs as shown by the broken line S in FIG.
1. , f 2 . . . Jn are shown as the correlation coefficient update time points.

Therefore, while driving on the road LO,
Until it is determined that the person has entered the road L1 through branch a (until the rotation angle data reaches a predetermined value or more), only the correlation coefficient 70 is large, and the other correlation coefficients 71 . γ2. γ3
is small (see region R1 in FIG. 3). In other words, until passing branch a, the rotation angle data is small, so there is no need to consider the influence of other roads at all, and only the travel distance on the road LO is a problem, so the estimated position calculation What is necessary is to correct the position data calculated in part (4) so that it becomes position data above the road LO. Then, by displaying the position along with the road map on a display device (not shown) based on the corrected position data, the current position of the vehicle can be confirmed.

Next, when entering road Ll through branch a, the correlation coefficient calculation unit (9 calculates the correlation coefficient γ for each road existing within the range of the above-mentioned limit error E).
(see region R2 in the figure) is evaluated by the correlation coefficient evaluation unit (6) based on the calculated correlation coefficient. That is, it is evaluated which change in estimated position has the highest similarity to the road. To explain in more detail, since the rotation angle data is equal to or greater than a predetermined value, the correlation coefficient 70 for the road LO can be ignored, and conversely, for roads where the branch for the road LO exists within the range of the limit error E. L 1. L 2. The correlation coefficient for L 3 must be evaluated. In FIG. 3, at the beginning of passing through branch a, the correlation coefficient is 72.
γ3 is larger than the correlation coefficient 71, but this is due to an error caused by the traveling direction detection unit (2) (However, in general, the correlation coefficient of the estimated position on the road Ll the relationship coefficient γl is the largest). However, regarding the correlation coefficient 72 of the estimated position on the road L2, since the direction detected by the driving direction detection unit (2) after passing the curve b and the direction of the road L2 are significantly different from each other, The correlation coefficient 73 of the estimated position on the road L3 decreases rapidly, and the direction detected by the driving direction detection unit (2) after passing through the branch C and the direction of the road L3 are significantly different. Because of the deviation, the correlation coefficient 71 of the estimated position on the road LL decreases rapidly.On the contrary, the correlation coefficient 71 of the estimated position on the road LL hardly changes. The estimated position above can be determined as the current position and displayed on a display device (not shown) together with the road map. That is, until reaching branch a, the current right position is displayed in correspondence with the top of road LO, and after passing branch a, until the estimated position on road L2 passes curve b. During this period, the current position is displayed corresponding to the top of the road L2, and the period from when the estimated position on the road L2 passes through curve b until the estimated position on the road L3 passes through branch C. The current position can be displayed in correspondence with the road L3, and after the estimated position on the road L3 passes the branch C, the current position can be displayed in correspondence with the road L1. .

In FIG. 3, when the estimated position on the road L1 is determined as the current position, the correlation coefficient γl is set to the maximum value (for example, 1.0).

Therefore, after that, by performing the same process again using the display position on the road L1 as a reference, it is possible to display the current position accurately as the vehicle travels.

Note that the accurate determination of the current position may be performed only in units of a predetermined unit of the road, for example, as shown by curve d in FIG. 3, or if the correlation coefficient γ is a predetermined threshold ( (not shown), the registration of the estimated position corresponding to the corresponding correlation coefficient may be deleted, and furthermore, when the difference between the correlation coefficients exceeds a predetermined threshold, the registration of the estimated position corresponding to the corresponding correlation coefficient may be deleted. The registration of the estimated position corresponding to the smaller correlation coefficient may be deleted.

In addition, in the case of the above example, the highest correlation coefficient changes from γ2 to γ3 to γ1, and the display on the display device also moves from above road L2 to above road L3 and then Ll, so However, in the case where the order of the correlation coefficients is reversed, by providing a predetermined hysteresis characteristic, it is possible to prevent the display on the display device from rapidly and frequently changing to different roads. .

Next, explanation will be given based on FIGS. 4 and 5.

Figure 4 is a diagram showing a part of the road transportation network, including roads LO,
Only LL and L2 are shown, and the vehicle is on the dashed line S.
As shown, the vehicle is traveling from road LO to road Ll. In addition, E indicates the marginal error, a,
b and c indicate branches or curves, and γ0. γ1. γ2 indicates a correlation coefficient corresponding to the estimated position on each road.

.

Moreover, FIG. 5 is a diagram showing the change in the correlation coefficient γ when the vehicle runs as shown by the broken line S in FIG.
1. J 2 . . . 1n is shown as the correlation coefficient update time point.

Therefore, while driving on the road LO,
Until it is determined that the person has entered road L-1 through branch a (until it is determined that the person has passed branch a based on the distance data), only the correlation coefficient 70 is large, and the other correlation coefficients are large. Relationship number 71. γ2 is small (see region R1 in FIG. 5). In other words, since the rotation angle data is small until passing branch a, there is no need to consider the influence of other roads, and road L
Since it is only necessary to consider the traveling distance at O, it is sufficient to correct the position data calculated by the estimated position calculation unit (4>) so that it becomes the position and data above the road LO. Based on the corrected position data, the current position of the vehicle can be confirmed by displaying the position along with the road map on a display device (not shown).

Next, when entering road L1 through branch a, the correlation coefficient calculation unit (calculate the to-to-eye relationship coefficient γ for each road Ll and L2 existing within the range of the above-mentioned limit error E) (see region R2 in FIG. 5), is evaluated by the correlation coefficient evaluation section (6) based on the calculated correlation coefficient. That is,
Evaluate which change in estimated position has the highest similarity to the road. To explain in more detail, since we have passed branch a, the correlation coefficient for the road LO is now 70.
can be ignored, and conversely, the correlation coefficient for roads Ll and L2 branched at branch a must be evaluated. In FIG. 5, the correlation coefficient γ1 is slightly larger than the correlation coefficient 72 at the beginning of passing through branch a, but both correlation coefficients 71. Since the difference in γ2 is very small, both roads are LL.

The estimated position above L2 is displayed on a display device (not shown). Of course, only the estimated position on Ll may be displayed on a display device (not shown). However, road L
For the correlation coefficient 72 of the estimated position above 2, curve b
After passing through the road L2, the deviation between the direction detected by the driving direction detection unit (a) and the direction of the road L2 becomes large.
Then, after passing through curve C, the direction detected by the driving direction detection unit (2) and the direction of road L2 are significantly different, so it suddenly becomes smaller.On the contrary, the estimated value on road L1 The positional correlation coefficient γl hardly changes. Therefore, based on the relatively large correlation coefficient 71, the estimated position on the road L1 is determined as the current position, and is displayed on a display device (not shown) together with the road map. In other words, until reaching branch a, the current position is displayed corresponding to the road LO, and after passing branch a, the estimated position on road L2 passes curve C. Until then, please use roads LL, L2 or
The current position is displayed in correspondence with only the top of Ll,
After the estimated position on the road L2 passes through the curve C, the current position can be displayed only on the road L1.

In FIG. 5, when the estimated position on the road L1 is determined as the current position, the correlation coefficient 71 is set to the maximum value (for example, 1.0).

Therefore, after that, by performing the same process again using the display position on the road L1 as a reference, it is possible to display the current position accurately as the vehicle travels.

The limit error E is different in both examples, and is larger in the examples shown in FIGS. 2 and 3 because the straight-line traveling distance is longer.

However, in the case of general driving in urban areas, etc., a branch or curve is often reached before the limit error E becomes too large, and there are two or more points within the limit error E in this state. It is rare for a road to exist, and even if two or more roads exist, there is a high probability that the correlation coefficient will be larger for the road that corresponds to the actual driving. As shown in , it is rare that a location on a different road is recognized as the current location.

As is clear from the above explanation, by outputting the current position inside the vehicle, it can be applied to a navigation system that displays the current position and destination together with a road map to provide route guidance. By outputting the position to the outside of the vehicle using radio waves and receiving the radio waves at one location, the present invention can also be applied to a location system that grasps the running status of a large number of vehicles.

Note that the present invention is not limited to the above-mentioned embodiments. For example, if there are multiple estimated positions with the highest correlation coefficient, the estimated position closest to the center of the road may be output as the current position. In addition to being possible, the mileage for each predetermined distance,
Instead of obtaining and integrating the amount of change in distance and direction, it is possible to obtain and integrate the amount of change in travel distance and direction at predetermined time intervals, and various other methods may be used without changing the gist of the present invention. It is possible to make design changes.

<Effects of the Invention> As described above, the present invention can be applied to all roads that exist within the limits of the error that a sensor for detecting the driving state of a vehicle inevitably has and the error of a road map. By registering location data and calculating correlation coefficients for all the roads mentioned above, the similarity of roads to actual driving patterns can be evaluated. Even if the road is complex or there are multiple roads in close proximity to each other, two or more estimated positions including the accurate current position are always maintained, even if the position is temporarily incorrect. Even if an estimated position that is not available is output as the current position, the correct estimated position can be output as the current position based on subsequent changes in the correlation coefficient.
This has the unique effect of significantly improving the precision and accuracy of position detection.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of a device for implementing the position detection method of the present invention, FIGS. 2 and 4 are diagrams showing a part of a road transportation network, respectively. and FIG. 5 are diagrams showing changes in correlation coefficients corresponding to FIGS. 2 and 4, respectively. (1)... Traveling distance detection section, (2... Traveling direction detection section, (3)... Road map memory, (4)... Estimated position calculation section, (S... Correlation coefficient Calculation unit, (6) Correlation coefficient evaluation unit Patent applicant Sumitomo Electric Industries Co., Ltd. Representative Patent attorney Hiroshi Kamei Figure 4 Figure 5 Mileage distance

Claims (1)

[Claims] 1. Obtain and integrate the distance traveled and the amount of change in direction at each predetermined timing, and based on the integrated data, determine your own position on any road on a road map stored in memory. In a self-contained position detection device for a vehicle, it obtains its own estimated position determined based on the accumulated data, the amount of error in the estimated position determined based on the error between the accumulated data and the road map, and calculates the error centered on the estimated position. The above estimated position is registered as the self-position in correspondence with all roads located within the range of the amount, and the correlation coefficient of the registered estimated position for each road is calculated to find out that the error for the road is the smallest. A position detection method characterized by selecting a correlation coefficient shown in the table and outputting an estimated position corresponding to the selected correlation coefficient as a current position. 2. The position detection method according to claim 1, which selects the largest correlation coefficient and outputs the estimated position corresponding to the largest correlation coefficient as the current position. 3. The position detection method according to claim 2, which cancels the registration of the corresponding estimated position when the correlation coefficient becomes equal to or less than a predetermined value. 4. The position detection method according to claim 2, which cancels the registration of the estimated position corresponding to the smaller correlation coefficient when the difference between the correlation coefficients exceeds a predetermined value. 5. The position detection method according to claim 3 or 4, wherein the registration of the estimated position is deleted after passing through a predetermined unit of the road. 6. In the above-mentioned claims, when the estimated position with the largest correlation coefficient and the estimated position with the second largest correlation coefficient are switched, the largest correlation coefficient is selected with a predetermined hysteresis characteristic. The position detection method described in Section 2. 7. When there are multiple estimated positions with the maximum correlation coefficient,
The position detection method according to claim 2, wherein the estimated position closest to the center of the road is output as the current position. 8. The position detection method according to claim 2, wherein the distance traveled and the amount of change in direction are obtained and integrated at predetermined time intervals. 9. The position detection method according to claim 2, which obtains and integrates the travel distance and the amount of change in direction every predetermined distance.